How to Choose a PhD Program

14 minute read


Hello and congrats! You’ve decided to pursue a Ph.D. It is an extremely exciting time, and I’m glad that you’ve decided to go to graduate school. It continues to be one of the best decisions of my life (this may seem extreme but really, it’s super fun).

However, applying to graduate school can be extremely overwhelming. Hopefully this post makes it less so. As a disclaimer, all of this is from my own experience. There are a number of other wonderful blogs and pages about applying to graduate school that are also helpful. In addition, most (if not all) of this information specifically pertains to applying to computer science PhD programs in the United States.

Where do I apply?

The first natural thing to do is figure out how many graduate programs you want to apply to. This number can vary a lot depending on your goals (whether you absolutely want to go or if you are considering industry options as well), the strength of your application, and your domestic/international status. What I have heard as general guidance from others is that if you are a domestic student, $8-12$ applications is a good range. If you are an international student, that number might be higher at around $15-25$ applications. In general, having a spread of “reach” schools, “target” schools, and ``safety” schools is a good idea. I personally applied to $17$.

A good thing to note is that graduate admissions are extremely different from undergraduate admissions. For example, similarly to undergraduate admissions, you usually are applying to one department (i.e. computer science, mathematics, physics). However, you are unofficially applying to a subgroup as well. I listed myself as a theory (algorithms and complexity) student on my applications. Furthermore, the number of available spots within this subgroup (so for me, the theory group) can heavily depend on the year. Factors include how many students the subgroup took in previous years and how much funding is available.

When determining where to apply, a good place to start is CS Rankings. I used this to get a sense of what the top $30$ theory programs were. Then, I immediately removed any schools that were located in deal-breaker locations. I also removed schools that seemed especially problematic. I personally struck a school off my list for having a large number of sexual assault cases. Then, for each school, I looked at the theory department and wrote down every single professor I would possibly be interested in working with.

To narrow down the schools from $30$ to the $17$ I ended up applying to, I relied heavily on the expertise of those around me. I was extremely lucky to have an undergraduate advisor who helped me pick good schools and potential advisors. If you know anyone in the subfield you are planning to go into, they are probably the best source of information. If that is not available, here are a few things to look for. For a particular school,

  • Are there a number of professors you could work with? Remember that there is a concrete possibility you’ll have to change advisors during your PhD.
  • What’s the size of the research group you’re interested in? In general, larger groups may mean many potential collaborators, but smaller groups may be more tight-knit.
  • Where are graduated students now working? Does the department place a lot of students in academia vs. industry?
  • Location?
  • Pay - although this probably won’t be the determining factor, it is a lot easier to live on the stipend at, say, CMU than it is at UC Berkeley. That can matter when you’re in a high stress PhD program.

For potential advisors,

  • Are their previous students successful?
  • Do they have a history of working with underrepresented students? For me, it was important that a potential advisor did not have a huge list of students with no women on it.
  • Do they publish good results with their students? Some professors are successful themselves and write amazing papers, but don’t produce the same quality of papers with their students.
  • Do their students work on a somewhat diverse set of problems? This can indicate how flexible an advisor is.
  • Does their research interest you?

How do I apply?

I won’t expand too much in this section, as there are tons of resources about the actual application to graduate school. I’ll just go over the highlights.

For computer science, you generally need a personal statement and at least $3$ letters of recommendation. A lot of schools also now require a diversity statement. While there are a number of schools that require/recommend the GRE, almost everyone I know has gotten away with applying to all the programs they want without taking it.

Your personal statement should usually be around $2$ pages. A lot of professors do not read the personal statement and make admission decisions almost entirely on your previous experience and recommendation letters. However, there are still some professors who do read it. Here are a few things that professors have told me about the personal statement.

  • It should demonstrate that you know how to do research. Namely, when talking about your previous research experiences: how did you approach defining the problem? How did you approach solving the problem and what did you learn from the process? This is usually better than a bullet point list of contributions that someone could find in the abstract of your papers.
  • It should include what you are generally interested in researching and why.
  • It should demonstrate your work ethic and desire to do research.
  • If you have been involved in service / the CS community, you should include those contributions.
  • Make sure to include the top $2-3$ professors that you are interested in. This greatly boosts the chance that they read your application.

Recommendation letters are often cited as one of the most crucial pieces of your application. $1-2$ letters of recommendation should come from someone who you worked closely with, preferably on a research project. That would be ideal for all $3$ letters but is unrealistic for most applicants. For my third letter, I asked a teacher whose research I thought was cool and whose class I had done well in.

I also want to note that a lot of schools have application waivers for groups that are traditionally underrepresented in computer science. See my post on resources for women in computing (coming soon) for a semi-comprehensive list (although you should just check for each school you’re interested in applying to).


I will just outline my own experiences and thoughts about interviewing. In general, I got interview requests around the end of January to the beginning of February. This timeline heavily depends on which school you’re applying to. A professor from the school would email me, and then we would set up a time for the interview within the next week or so. From my experience, interviews were pretty informal. At most schools, professors simply pick a few students they might be interested in working with to interview. For some others (including CMU), subgroups interview as a group.

As for the actual interview, most of mine lasted around an hour. It started off with the professor asking me a few questions about my past research experience. Then, they would talk about the research they are interested in and what working with them might look like. At the end, there was usually some time for me to ask questions.

Interviews are a way for professors to vet that you are who you say you are. In addition, it’s a way for them to see if you might be a good fit as a student. Although it may not seem like it, interviews are also highly beneficial to you! First, you can vibe-check the interviewer. In general, this interviewer is likely the professor interested in you at this particular school. So, if you went to the school, there’s a good chance that you would work with them.

Also, it’s a great opportunity to talk to and learn from a highly established researcher. A lot of these professors are very successful and very busy. This means that it can be hard to find opportunities to talk to them, especially one-on-one. So, this is a great time to get to talk about your research with a super cool, super smart person. And, if you are interested in their research, you get to ask this person questions too! All in all, interviews were a bit stressful but still extremely fun.

I want to stress that while a lot of schools interview now due to the convenience of Zoom, there are a lot of schools and professors who admit without interviewing. So, if you don’t get an interview request from a place, that doesn’t mean that you should count on receiving a rejection.

Choosing a school

Congratulations! You’ve gotten in. What now?

The fun part about this section of the timeline is schools and professors are recruiting you! Savor it - it’s a time when many people who are very busy will take time out of their schedules to speak with you. And, it feels nice to be chased :). I’ll outline some of the things that helped me decide where to attend. Many of these are things you can find out by participating in visit days (which I highly recommend).

These are a few things I looked at regarding the student body at a particular school.

  • Are there other students who work on what you want to work on? Yes means more potential collaborators and people to help you!
  • Are the other students generally welcoming and kind? Or do they seem overly competitive?
  • Do the current students seem happy in the program? Usually they will be pretty honest about the pros/cons of the program if you ask them.
  • Do the current students hold events, social and/or educational? (outings, lunches, seminars, reading groups)

Here are some other things to consider about a particular school. I personally found these less important.

  • Does the administration seem responsive and helpful?
  • Is the location and campus nice?
  • Are there common spaces for collaboration?

Here are some things that I looked for when considering potential advisors.

  • Will they take you as a student? This is SUPER important. It is completely acceptable (and even recommended) to ask professors you’re interested in if they would be able to work with you.
  • Do they have other students? Are those other students sociable? Do their other students collaborate with each other?
  • Is funding tight? This can be important for things such as summer funding, how much you need to TA, conference funding, etc.
  • Do they seem like a nice person that you would enjoy working with for $5-7$ years of your life?

One of the most important things to me when choosing a graduate school was having an advisor who was a good fit. I luckily found two co-advisors who have been wonderful. Here are the questions that I asked each professor I was interested in and their students to make sure I encountered no surprises.

To the professors themselves (in no particular order):

  • If I came in the fall, would you be able to take me as a student? Please, please, please ask this. Even if just to make sure.
  • How do you feel about co-advising?
  • Would you consider yourself a hands-on or hands-off professor? Would you work on the technical aspects of projects with me?
  • How often would we meet?
  • How flexible are you with what problems we would work on?
  • How often do your students tend to collaborate with outside researchers?
  • How much travel are you okay with me doing? Are you okay with remote work? This can be especially important for international students who want to go home in the summers.

These questions may seem a bit blunt. I would advise you to ask them in an extremely polite way. However, they are super informative questions that will help you understand who this person is and what working with them might look like. Of course, there is nothing to replace actually working with them, but you can get decently close. In my opinion, any professor who won’t give you relatively honest and direct answers to these questions may not be someone you want to work with.

To their students:

  • What do you like about your advisor? What do you dislike?
  • Do they give you feedback on how you’re doing?
  • Do they help you set up meaningful collaborations?
  • Do you feel like they are invested in you and your growth?
  • How much flexibility do you have with what problems you want to work on?

Ultimately for me, I chose CMU because of the large and happy student group and my now advisors. I chose my advisors partly because of research interest but also largely because of advising style. Going into graduate school, I wanted hands-on advisors that were nice people. So, that’s what I looked for.

A note of warning. There are a number of professors who despite being extremely successful in their careers, are known to be bad advisors. You will likely hear about such professors during your visit days. It is generally good to stay away from those professors; students are not exaggerating their claims about them.

I hope this helped! Congratulations again! Although I’ve only just finished my first year, graduate school has been one of the best experiences of my life. You should feel extremely proud and excited.